Advancements in computing technology have led to the production of sensors capable of observing and reporting various real-world phenomena in a time-sensitive manner. Additionally, the growth in distributed communication technology (e.g., the Internet) has led to the development of sensor networks. Sensor networks have been proposed for use in numerous applications, including military and civilian applications. Generally, sensors are adapted to detect or monitor certain events or conditions. Sensors may be simple, such as a device that monitors temperature, or more complex, such as a video camera. Data generated at the sensor is transmitted in data packets over a sensor network to one or more end-points. An end-point can include an application software instantiation that can react to the sensor data or can be a user interface that presents the sensor data in numerical or graphical form to a user. Network traffic comprising sensor data are referred to herein as a sensor data flow. As the number of sensors increases, the probability of congestion in the sensor data flow increases which can lead to sub-optimal routing performance. Data packets can be dropped and the overall response time of the application or the user can increase.
Data aggregation is a technique known in the art that attempts to alleviate localized congestion problems. Generally, data aggregation is used to determine what data are useful and then to transmit only the useful data to the end-point, thereby reducing congestion and other associated problems. Various aggregation techniques have been proposed in the art. Although current data aggregation techniques have resulted in reduced congestion in sensor data flows, there is still a need for improved control of sensor data flows. As an increasing number of sensors get deployed over increasingly wider geographies and are networked to sets of applications over different access and IP networks, localization aspects, Quality of Service (“QoS”) aspects, and the relationship between the kind of event detected or condition monitored and an appropriate response to the event or condition become increasingly difficult to maintain.